Limitations of cytochrome oxidase I for the barcoding of Neritidae (Mollusca: Gastropoda) as revealed by Bayesian analysis S.Y. Chee Center for Marine and Coastal Studies, Universiti Sains Malaysia, Penang, Malaysia Corresponding author: S.Y. Chee E-mail: [email protected] Genet. Mol. Res. 14 (2): 5677-5684 (2015) Received September 5, 2014 Accepted February 26, 2015 Published May 25, 2015 DOI http://dx.doi.org/10.4238/2015.May.25.20 ABSTRACT. The mitochondrial DNA (mtDNA) cytochrome oxidase I (COI) gene has been universally and successfully utilized as a barcoding gene, mainly because it can be amplified easily, applied across a wide range of taxa, and results can be obtained cheaply and quickly. However, in rare cases, the gene can fail to distinguish between species, particularly when exposed to highly sensitive methods of data analysis, such as the Bayesian method, or when taxa have undergone introgressive hybridization, over-splitting, or incomplete lineage sorting. Such cases require the use of alternative markers, and nuclear DNA markers are commonly used. In this study, a dendrogram produced by Bayesian analysis of an mtDNA COI dataset was compared with that of a nuclear DNA ATPS-α dataset, in order to evaluate the efficiency of COI in barcoding Malaysian nerites (Neritidae). In the COI dendrogram, most of the species were in individual clusters, except for two species: Nerita chamaeleon and N. histrio. These two species were placed in the same subcluster, whereas in the ATPS-α dendrogram they were in their own subclusters. Analysis of the ATPS-α gene also placed the two genera of nerites (Nerita and Neritina) in separate clusters, whereas COI gene Genetics and Molecular Research 14 (4): 5677-5684 (2014) ©FUNPEC-RP www.funpecrp.com.br S.Y. Chee 5678 analysis placed both genera in the same cluster. Therefore, in the case of the Neritidae, the ATPS-α gene is a better barcoding gene than the COI gene. Key words: Bayesian analysis; Cytochrome oxidase I; ATPS-α; Molecular taxonomy; Neritidae INTRODUCTION DNA barcoding is a technique used to characterize specimens of organisms using a short DNA sequence from a standard position in the genome. The mitochondrial DNA (mtD- NA) gene cytochrome oxidase I (COI) is becoming the standard barcode region for higher animals, as it is a relatively short sequence (648 nucleotide base pairs in most groups) and can be obtained reasonably quickly and cheaply. However, there are limitations of COI as a barcoding gene. The first reservation is its inability to distinguish between species if introgres- sive hybridization, over-splitting, or incomplete lineage sorting occurs (Steinke et al., 2009). Secondly, the COI gene is not sensitive enough to discriminate between species that have re- cently evolved, and thirdly, if the species in question exhibits extensive spatial differentiation, discovering new species is difficult (Hickerson et al., 2006; Rubinoff et al., 2006). In addition, COI gene amplification is not always successful for species-level identification. Some studies, including Meier et al. (2006) and Johnson et al. (2008), have not been completely successful in identifying species using DNA barcoding methods that involved COI. In studies such as Pleijel et al. (2008), which used the same method, specimens were misidentified. In such cases alternative markers are required, such as nuclear DNA (nDNA) mark- ers, as they are less sensitive to introgression or incomplete lineage sorting and are more ap- propriate for recently diverged species (Baker et al., 2009; Zhang, 2011). According to April et al. (2011), for morphologically distinct species that remain genetically indistinguishable, it is highly unlikely that they would be delineated, as they could be derived from the same gene pool. Indeed, nuclear genes have been used in identifying indistinguishable species of freshwater fish (April et al., 2011) when COI could not delineate them. Nuclear genes have also been used with mitochondrial genes to provide accurate genetic identification and over- come the problem of misinterpretation caused by hybridization and introgression in museum samples of sturgeons (Garrido-Ramos et al., 2009). In the barcoding and identification of 17 known and easily confused species of Muricidae species from China, nuclear genes, and COI, were used to offer an analytically powerful addition or even an alternative (Zou et al., 2012). The specimens used in the present study were nerites (Neritidae) from the Proso- branchia family. Nerites are operculate sea snails with a short spiral shell. They are widely distributed in tropical and temperate coastal waters. Some species are found on rocky and/ or sandy shores, whereas others can only be found in muddy mangroves, and are commonly spotted hiding in crevices to protect themselves from predators and physical stress. They are gregarious herbivores, and play important roles in the food web and as bioindicators. Some of the more abundant species are important to algal community structure in coastal areas. Nerites are abundant, easy to collect, hardy, and relatively long-lived. To date, no studies concerning the species identification of Malaysian nerites have been conducted; the paucity of gastropod taxonomists has also hampered the process of determining their biodiversity. This study at- tempted to identify nerites found in Malaysian intertidal zones, using the barcoding mtDNA Genetics and Molecular Research 14 (4): 5677-5684 (2014) ©FUNPEC-RP www.funpecrp.com.br Limitations of COI for barcoding nerites 5679 COI gene and the nDNA ATPS-α gene. Bayesian data analysis was employed in this study. Bayesian refers to methods in probability and statistics, and is considered to be fundamentally sound, very flexible, produces clear and direct inferences, and makes use of all available information (Pernestal, 2009). This method has been used to provide accurate estimations of the evolutionary rates, morphologi- cal parallelism, and biogeography of the Littorininae (Williams et al., 2003); phylogenetic relationships and speciation in calyptraeid gastropods (Collin, 2003); and the phylogeny and taxonomy of topshells (Trochidae) (Donald et al., 2005). We tested the sensitivity of COI for barcoding using this established mode of data analysis. MATERIAL AND METHODS Sampling Samples were collected from selected beach areas along the coasts of the Malaysian Peninsular, Sarawak, and Sabah, providing a complete geographical coverage throughout the distribution of nerites. In total, 14 locations were sampled. Samples of each morphospecies were collected from every location they could be found to avoid overlooking cryptic species. When possible, at least five individuals per morphospecies were collected per site. Photo- graphs of dorsal and ventral views were taken of each species to aid morphological descrip- tions and to catalogue the diversity of the Neritidae. Dissection was conducted in the field whenever possible, or in the laboratory, after the muscles were relaxed using 10% MgCl2 diluted in double-distilled water (ddH2O). The foot tissues of each specimen were then cut and placed into 1.5-mL tubes filled with absolute ethanol (99%), and subsequently stored at 27°C. Molecular analysis DNA was extracted from 2 mm3 of foot tissue using an AquaGenomic™ kit, following the protocol provided by the manufacturer (MultiTarget Pharmaceuticals LLC, Colorado Springs, CO, USA). The quality and quantity of the DNA extracted was checked at two wavelengths (260 and 280 nm) using a spectrophotometer (U-1900, Hitachi). Each DNA stock sample was diluted to a concentration of 50 ng. All DNA samples were then stored in a refrigerator at -2°C. The COI gene was amplified with the following primers: HCO2198 5'-TAA ACT TCA GGG TGA CCA AAA AAT CA-3' and LCO1490 5'-GGT CAA ATC ATA AAG ATA TTG G-3' (Folmer et al., 1994). Each polymerase chain reaction (PCR) contained 2.5 µL 10X PCR buf- fer, 2.5 µL 25 mM MgCl2, 0.5 µL 2.5 mM dNTPs, 0.25 µL 10 µM forward and reverse prim- ers, 0.1 µL 5 U Taq polymerase, and 1.0 µL 50 ng DNA template. ddH2O was added to make a total volume of 25 µL. Amplification consisted of 40 cycles of denaturation at 94°C for 30 s, annealing at 45°C for 1 min, and extension at 72°C for 1 min. The nuclear ATPS-α gene was amplified using the following primers: ATPSαf1 5'- GAG CCM ATG CAG ACT GGT ATT AAG GCY GT-3' and ATPSαr1 5'-TTG AAN CKC TTC TGG TTG ATG GTG TC-3' (Jarman et al., 2002). Each PCR contained 2.5 µL 10X PCR buffer, 2.5 µL 25 mM MgCl2, 1.0 µL 2.5 mM dNTPs, 1.0 µL 10 µM forward and reverse prim- ers, 0.2 µL 5 U Taq polymerase, and 1.0 µL 50 ng DNA template. ddH2O was added to make a total volume of 25 µL. The PCR amplifications consisted of an initial denaturation at 94°C for 5 min, 35 cycles of denaturation at 94°C for 45 s, annealing at 60°C for 45 s, extension at Genetics and Molecular Research 14 (4): 5677-5684 (2014) ©FUNPEC-RP www.funpecrp.com.br S.Y. Chee 5680 72°C for 1 min, and a final extension at 72°C for 7 min. All of the PCRs were conducted in a G-Storm™ Thermal Cycler GS4822 (Somerset, United Kingdom). Prior to sequencing, the PCR products were purified using a Wizard® Genomic DNA Purification Kit (Promega) to remove residual primers and dNTPs. The purified products were then sent for sequencing at the Universiti Sains Malaysia Center for Chemical Biology facility and run on a 3130 Capillary Electrophoresis Genetic Analyzer (Applied Biosystems Inc.), fol- lowing the manufacturer protocol. Unique sequences were deposited in GenBank and BOLD systems (version 2.5). Data analysis Sequences were edited and trimmed using MEGA (version 5.10) (Tamura et al., 2007), aligned with ClustalW, and checked for accuracy by eye in Chromas (version 2.33) (Technelysium, 2009).